Austen Lamacraft - University of Virginia · 2009-04-13 · Magnetism in condensed matter Magnetism is one of the pillars of condensed matter –One of the most studied solid states

faculty.virginia.edu/austen/

Magnetism in ultracold gases

Austen Austen LamacraftLamacraftTheoreticalTheoretical condensed matter and atomic physics condensed matter and atomic physics

April 10th, 2009April 10th, 2009

Outline

Magnetism in condensed matter

Ultracold atomic physics

Magnetism in Bose gases

Spontaneous symmetry breaking

Magnetism in condensed matter

Magnetism is one of the pillars of condensed matter– One of the most studied solid states of matter

(along with metals, superconductors, insulators, etc.)– Huge technological importance e.g. magnetic storage– A quantum phenomenon - depends on electron spin

Thales of Miletus

Magnetism comes in many flavors

Plain vanilla magnet is Ferromagnet (like Fe!)– Electron spins line up (exchange interaction from Pauli)

Can also have Antiferromagnetism (e.g. MnO)

Basic research in magnetism

Asks the question: is magnetic order inevitable?(at low temperatures)

Not all materials freeze at T=0– Helium remains liquid due to quantum fluctuations

Can same happen to magnetic moments of spins?– Spin liquid - hypothetical state of matter

Ultracold atomic gases

Fe becomes ferromagnetic at T=1043 K Ultracold atomic physics takes place at <10-6 K

Quantum effects determine collective (i.e. material)properties when

thermal wavelength ≈ interparticle separation

Atomic gases are heavier and less dense than gasof electrons in FeVast differences in scale no obstacle to theorists!

So what’s new?

In the solid state we (mostly) care about the quantummechanics of electrons. These are fermions

By contrast, atoms (considered as particles) may bebosons or fermions

Possibility of Bose-Einstein condensation - bosonsaccumulate in lowest energy state

– Nobel prize 2001

Exotic magnetism

87Rb has nuclear spin I=3/2, electron spin S=1/2– Possible total spin F=1 or 2

What are magnetic properties of F=1 or 2 Bose gas?

Magnetism in Bose gases

BEC: (nearly) all atoms sit in same quantum state This state is called the condensate wavefunction But what if lowest energy state is degenerate?

Condensate wavefunction is a spin vector (spinor) andmust pick a direction in spin space

Bose condensates with spin are always magnets

Why higher spin is fun

Spin 1/2 (e.g. of electron) points in some direction

– To make electron magnetism more interesting need non-trivial arrangements on lattice (e.g. antiferromagnetism)

Spin 1 doesn’t necessarily “point” anywhere

– Fi spin-1 matrices– Yet evidently there is still an axis involved!

Which spin state wins?

Must consider interatomic interactions Atoms can collide with total spin 0 or 2

– Total spin 1? Antisymmetric and blocked by Bose statistics

– Spin 2

– Spin 0

Spin dependent interactions

Energy of state includes a piece

– For cF<0 (e.g. 87Rb): maximize <F> Ferromagnet– For cF>0 (e.g.23Na): minimize <F> Polar state

The Bose ferromagnet: 87Rb

– Stamper-Kurn group, Berkeley

Polar condensates and a paradox

For cF>0 minimize <F>. Pick a quantization (z) axis

Problem: for a more general state

seek ground state ofMust be a singlet

The parable of the chair

Chair has (rotational) Hamiltonian

States very unlike the chair we see Tiny energy differences swamped by perturbations

Symmetry breaking in atomic gases

The same goes for

From excitations on top of singlet ground state, wecan build a state

at little cost, with a definite axis (but still <F>=0)

Spontaneous symmetry breaking